(a) Field of the Invention
The present invention relates to an image processing method, and more particularly, to an image motion detecting method.
(b) Description of the Related Art
In general, image data are transmitted via a composite signal. The composite signal includes luminance information and chrominance information where the chrominance information is appended on to the luminance information through modulating the high frequency part of the luminance information.
Therefore, after receiving the composite signal, a video decoder has to separate out the luminance information and the chrominance information for subsequent image processing operations. Such separating operation is called luminance/chrominance separation (Y/C separation).
As is known by those who are skilled in the art, the Y/C separation can be classified into three categories. They are the 2D Y/C separation, the 3D Y/C separation, and the combinational 2D and 3D Y/C separation that lies in between the above two methods.
As the name implies, the 2D Y/C separation is performed by referencing to the 2D information. In other words, the 2D Y/C separation is performed according to the relation between each pixel and its neighboring pixels in the same picture (frame). The 2D Y/C separation is generally utilized in a motion image. This is because differences exist between any two pictures in the motion images. Therefore, if the corresponding pixels of different pictures are utilized as references for the Y/C separation, the processed images will have distortions.
The 3D Y/C separation is performed by referencing to the 3D information. In other words, the 3D Y/C separation is performed according to the mutual relationships between the corresponding pixels in different pictures (frames). The 3D Y/C separation is generally utilized in a still image. This is because the corresponding pixels in different pictures of a still image basically have the same pixel values, such as the luminance value and chrominance value. Therefore, a simple 3D filtering operation (for example: an averaging operation) can be performed to recover the luminance value and the chrominance value correctly.
And, another separating operation that lies in between the 2D and the 3D Y/C separating operations is the combinational 2D and 3D Y/C separation. Such operation determines the weights (allocated proportion) of the 2D Y/C separation and the 3D Y/C separation according to the motion condition (such as: the extent of still) of an image. Thus, the Y/C separation can also be achieved by the combinational separating operation.
From the above descriptions, it can be seen that those who are skilled in the art need to use different Y/C separating operation to process an image according to the property of the image such that the image quality could be maintained. In other words, before an appropriate category of the above-mentioned three Y/C separations is selected, the property of the image (that is, the above-mentioned image motion condition) should be detected first. Then, an appropriate Y/C separating operation can be selected according to the motion condition of the image after the motion condition of the image is determined.
However, it is not easy to correctly determine the motion condition of the image. For example, because the image signal is often interfered by noises, it is sometimes hard to distinguish whether a detected high frequency phenomenon comes from the property of the image itself or the noise interferences while performing motion detection on the image. From the above description, it can be seen that the selection of the Y/C separation method is directly influenced such that the quality of the processed image becomes worse when the motion condition of the image is incorrectly determined. For example, if the still portion of an image is determined as having motions, the 2D Y/C separating operation is thus selected to process the image. Then, the boundary of an object or the line edges shown in the image may be blurred and the cross-color phenomenon may occur.
Therefore, the image motion detecting method becomes an important topic of the image processing method.